The ALOS PALSAR Global Systematic Acquisition Strategy:
4 Years in Operation
A. Rosenqvist a, M. Shimada b, T. Tadono b, K. Tsuzuku c, T. Watanabe c, E. Aoki c and F. Ohgushi c
a
soloEO-Japan, TTT Mid-Tower 1708, Kachidoki 6-3-2, Chuo.ku, Tokyo 104-0054, Japan – ake.rosenqvist@soloEO.com
b
JAXA EORC, Sengen 2-1-1, Tsukuba-shi, Ibaraki 305-8505, Japan – (shimada.masanobu, tadono.takeo)@jaxa.jp
c
RESTEC, Sengen 2-1-1, Tsukuba-shi, Ibaraki 305-8505, Japan – (tsuzuku_kaoru, tomozo, aoki_emi, fumi_og)@restec.or.jp
Abstract — With the launch of ALOS in 2006, JAXA took
the initiative to implement the first global-scale systematic
acquisition strategy for satellite sensors at fine (2.5m-20m)
spatial resolution. Comprising all three sensors on ALOS
(PALSAR, PRISM, AVNIR-2), the plan is designed to serve
all ALOS user categories and aims at producing spatially
and temporally consistent coverages over the planet on a
repetitive basis, to accommodate systematic global-scale,
fine-resolution, monitoring of the environment. Unlike the
common background missions defined for most fineresolution Earth Observation satellites, the BOS has been
implemented as a top-level foreground mission with a
priority second only to that of special observation requests
and emergency observations and sensor cal/val.
Presently 4 years in operation, the strategy has produced a
comprehensive and homogeneous global archive in which
consistent time-series of data are available for any arbitrary
point or region on Earth. Clouds and haze inevitably
constitute limitations for the oprtical sensors, while for the
PALSAR instrument, two cloud-free and near-gap free
(~95%) global coverages are achieved annually. Previously,
such uniform data archives existed only for coarse and midresolution sensors such as AVHRR, MERIS and MODIS.
Keywords:
strategy
ALOS PALSAR, BOS, systematic acquisition
1.
ALOS AND THE "BOS"
1.1 ALOS
The Advanced Land Observing Satellite (ALOS) was launched
by the Japan Aerospace Exploration Agency (JAXA) on
January 24, 2006. It carries three remote sensing instruments:
the along-track 2.5 metre resolution Panchromatic Remotesensing Instrument for Stereo Mapping (PRISM), the 10-metre
resolution Advanced Visible and Near-Infrared Radiometer
type 2 (AVNIR-2), and the variable resolution polarimetric
Phased Array L-band Synthetic Aperture Radar (PALSAR).
Following a 9-month comissioning phase, ALOS was declared
operational on October 23, 2006.
1.2 The Basic Observation Scenario (BOS)
Apart from the technical characteristics of the satellite and its
sensors, what makes the ALOS mission unique is the
acquisition concept, which differs significantly from the
"traditional" way fine resolution satellites are operated. The
acquisition plan, which is referred to as the "Basic Observation
Scenario" (BOS) by JAXA, is a fixed multi-year plan that was
developed prior to the launch of ALOS in support of the science
goals for the ALOS mission [1, 2]. It was developed in
collaboration with an international team of experts–the Kyoto &
Carbon Initiative Science Panel–and was designed to produce
spatially and temporally consistent coverages of the planet on a
repetitive basis, to accommodate systematic global-scale, fineresolution, monitoring of the environment. Unlike the common
background missions defined for most Earth Observation
satellites, the ALOS BOS has been implemented as a top-level
foreground mission with a priority second only to that of
emergency observations.
1.3 The PALSAR BOS characteristics
The Basic Observation Scenario has been implementd for all
three sensors on ALOS, but is particularly effectual for the
PALSAR sensor, which has the great advantage of not being
constrained by clouds or haze. The PALSAR BOS follows the
recommendations for systematic acquisitions outlined in [3] to
fulfil the following general acquisition concepts:
•
Spatial consistency: Continuous wall-to-wall
acquisitions over continental areas.
•
Temporal consistency: Each regional acquisition
performed during limited time windows to maximise
seasonal homogeneity in each wall-to-wall coverage
•
Revisit frequency: Revisit frequency adapted to meet
the requirements for the kay applications (semiannual repetition for forestry and land cover; everycycle repitition over inundated wetlands).
•
Timing: Acquisitions performed during the same
time period(s) every year to minimize temporal bias.
•
Sensor consistency: Selection of a limited number of
operational default modes to maximize data
homogeneity and minimise programming conflicts.
•
Mission-term continuity: Strategy operations
troughout the entire ALOS mission life.
The ALOS BOS is refined on a continuous basis and the latest
plan updates are available for viewing at the JAXA EORC
homepage: http://www.eorc.jaxa.jp/ALOS/en/obs/overview.htm
To limit the detrimental effect fragmented acquistions due to
mode changes and to assure spatio-temporal homogeneity over
regional scales, PALSAR features an observation plan in which
operations have been confined to a limited number of
operational modes. Out of the 132 original mode options
technically available for PALSAR, only five have been
identified as the main default modes (Table A).
To reduce programming conflicts further, acquisitions are
planned in units of whole (46-day) repeat cycles, during which
only one of the available default modes is selected.
Observations are also separated into descending and ascending
passes, with PALSAR operations first and foremost planned for
the latter to avoid conflicts with the optical sensors. As PRISM
and AVNIR-2 are confined to daytime operations, they are
assigned a higher observation priority than PALSAR, which, in
descending mode, principally is limited to low data rate (120
Mbps) ScanSAR observations.
Table A. ALOS PALSAR operational modes
Mode (polarisations)
FBS - Single polarisation (HH)
FBD - Dual polarisation (HH+HV)
POL - Quad-pol (HH+HV+VH+VV)
POL - Quad-pol (HH+HV+VH+VV)
ScanSAR (HH)
Offnadir
angle
34.3°
34.3°
21.5°
23.1°
5-beam
Incidence
range
Swath
width
Resolution
(4 looks)
Pass
designation
Coverage
36.6°~40.9°
36.6°~40.9°
22.8°~25.2°
22.8°~25.2°
18.0°~43.0°
70 km
70 km
30 km
30 km
350 km
10 m
20 m
~30 m
~30 m
~100 m
Ascending
Ascending
Ascending
Ascending
Descending
Global
Global
Semi-continental
Semi-continental
Regional/Global
As a consequence of the high data rate, neither the PALSAR
instrument nor PRISM or AVNIR-2 can achieve a full global
coverage within a single 46-day cycle, and observations have to
be divided regionally over several cycles. In the observation
plan, the Earth has been divided into some 80 adjacent, nonoverlapping geographical polygons which cover all land areas
and coastal regions (the same polygons are used for PRISM and
AVNIR-2), All passes withion each polygon are then acquired
to assure spatail and temporal consistency over regional scales.
Figure 1 shows the acquisitions for PALSAR during two
consecutive 46-day cycles in 2010, in which the geographical
stratification of the plan can be seen.
To assure recurrent observations with consistent timing, the
PALSAR observations are planned in groups of 8 (46-day) unit
cycles that are repeated on an annual (368 days) basis.
The PALSAR acquisitions comprise four global dual-season
coverages per year, two dual-polarisation (FBD) coverages in
May/June-Sept/Oct and two global single-polarisation (FBS)
coverages in Dec–March. Once every second year regional
acquisitions campaigns (two cycles in March-May) in full
polarimetric (POL) modes are undertaken. The Fine Beam
(FBD, FBS and POL) acquisitions are all designated to
ascending (night time) passes.
ScanSAR are planned to produce one full global coverage per
year, as well as every-cycle acquisitions over significant
wetland and forst change regions (South America, Central
Africa, SE-Asia). To avoid conflicts with the Fine Beam
acquisitions, ScanSAR acquisitions are all descending.
2. ACQUISITION RESULTS
The performance of the ALOS acquisition strategy during the
first four years of operations (Dec 2006 - Dec 2010) is shown
below in Table B.
The figures represent the average
acquisition success rate (ratio of #programmed scenes vs.
#acquired scenes)
for a 46-day cycle within the year
indicated.
Table B. ALOS acquisition success rate (ratio of
#programmed scenes vs. #acquired scenes) and number of
scenes acquired
for the first 4 years of operations
(cycles 7-39).
Figure 1. PALSAR acquisitions during two consecutive
cycles during the mid-2010 dual polariation (FBD)
acquistion period. Green–passes successfully acquired;
black–passes programmed, but not acquired.
The table shows that the average success rates for the two
PALSAR "work horse" modes–FBS and FBD– are in the order
of 78% for a given cycle, with some minor variations between
years. As gaps in the coverages always can be expected, the
BOS was designed to cover each geographical region two times
during each seasonal time window as a means of automated
gap-filling. While the distribution of missed acquisitions is not
entirely random and certain regions seem have a lower success
rate (e.g. central Asia and Siberia in Fig. 1), approximately 95%
complete coverage can generally be achieved over the 3-4
cycles of the seasonal FBS and FBD acquisition windows.
Figure 2. ALOS PALSAR coverage acquired during the mid-2009 observation window (3 cycles, June-October).
10 m resolution dual-polarisation (HH+HV) PALSAR mosaic (JAXA, 2010).
Colour composition: R–HH, G–HV, B–HH/HV ratio
An example of the above is the 2007 coverage over Africa,
where 319 path segments were required to cover the whole
continent. 276 of the passes were acquired during the first two
46-day cycles in June-August, 2007, another 27 passes during
the next 46-day cycle (Sept-Oct, 2007), while 16 passes (i.e. 5%
of the total) had to be filled in from the 2008 acquisitions.
Figure 2 shows a global mosaic generated by JAXA EORC
from PALSAR FBD data acquired during 3 cycles in the period
June to October, 2009 [4]. Such mosaics will be generated by
JAXA on an annual basis from the year 2007 and onwards.
3. CONCLUSIONS
2.1 Future plans
Minor modifications to the BOS are incorporated by JAXA on a
continuous basis and the latest schedules for the next few 46day cycles are made available on the JAXA EORC web:
http://www.eorc.jaxa.jp/ALOS/en/obs/overview.htm
The PALSAR BOS acquisitions are foreseen to continue
without major changes, with the exception that the annual FBS
acquisitions in December–March may be replaced by dualpolarisation FBD, to better accommodate dual-seasonal
monitoring of the global forest cover.
A systematic acquisition strategy that builds on the PALSAR
BOS is presently under development for the PALSAR-2
instrument on-board ALOS-2. ALOS-2 is an SAR-only satellite
(L-band of course) that is scheduled for launch in mid 2013.
2.2 GEO FCT & CEOS coordinated acquisitions
The need for systematic satellite acquisitions to support
monitoring of forest carbon in the UNFCCC post-2012 era was
raised with the Group of Earth Observations (GEO) at the 2nd
GEOSS Asia-Pacific Symposium in 2008, and the concept of
coordinated multi-sensor acquisitions was proposed as part of a
new GEO Task [5, 6]. The GEO Forest Carbon Tracking Task
(GEO-FCT) was subsequently established [7] and several space
agencies joined forces in 2009 to undertake the very first
coordinated multi-satellite acquisition campaign over national
scales involving optical, L-band, C-band and X-band SAR
sensors. The GEO-FCT observation plan was developed with
the ALOS global acquisition strategy as the model, featuring
semi-annual wall-to-wall acquisitions during common dualseason time windows (Dec-Feb and June-Aug), and with fixed
sensor modes to assure spatial and temporal consistency.
The GEO-FCT acquisitions are coordinated by the Committee
of Earth Observation Satellites (CEOS) and is lead by JAXA,
ESA, CSA and USGS. GEO-FCT is planned to develop into an
operational Global Forest Observation Initiative (GFOI),
featuring a coordinated observation strategy for all forest areas
on the globe.
REFERENCES
[1] Rosenqvist A., Shimada M., Ito N. and Watanabe M.
“ALOS PALSAR: A pathfinder mission for global-scale
monitoring of the environment,” IEEE Trans. Geosci. Remote
Sens, vol. 45, no. 11, pp. 3307-3316, 2007.
[2] Rosenqvist, A., Milne A.K. and Zimmermann, R., 2003.
Systematic Data Acquisitions - A Pre-requisite for Meaningful
Biophysical Parameter Retrieval?. IEEE Transactions on
Geoscience and Remote Sensing, Communication, Vol. 41, No.
7, pp.1709-1711
[3] Rosenqvist A., M. Shimada, M. Watanabe, T. Tadono and
K. Yamauchi, 2004. Implementation of Systematic Data
Observation Strategies for ALOS PALSAR, PRISM and AVNIR2. International Geoscience and Remote Sensing Symposium
(IGARSS’04). Anchorage, USA, Sept. 20-24, 2004.
[4] M. Shimada et al., “Generation of Global Forest/Non-Forest
Maps using ALOS PALSAR", JAXA EORC website: http://
www.eorc.jaxa.jp/ALOS/en/guide/forestmap_oct2010.htm
[5] Group on Earth Observations. Summary Report of 2nd
GEOSS Asia-Pacific Symp., Tokyo, Japan, April 14-16, 2008.
[6] Rosenqvist A., "Components of a systematic data
acquisition strategy". Proc. of the 2nd GEOSS Asia-Pacific
Symposium, Tokyo, Japan, April 14-16, 2008.
[7] The GEO FCT Portai: http://www.geo-fct